Yeah I generally agree with you. My view is that mobile will follow the desktop in terms of apps. Once upon a time, most apps were native to the desktop. Then the networks got fast enough and the browsers and tools got good enough that you could build most native desktop apps as websites. Once the mobile networks get fast enough, and the mobile browsers and tools get a bit better then I think you will start seeing mobile apps being built as websites too.

The main problem websites solve over native is the fact you only have to build the app once and it works everywhere (mostly). Of course, there will always be a place for native apps, regardless of desktop or mobile, when performance matters most.

I think you are already starting to see it happen actually with mobile-first web frameworks like Bootstrap, some mobile browsers performing well (i.e. Chrome), and 4G rolling out. But it is definitely just starting assuming I am right.

Which apps exactly, aside from shopping and maybe encyclopaedias, have migrated to the web ?

Desktop publishing ? Nope

IDEs ? Nope

CAD ? Nope

3D modeling ? Nope

Mathematical tools ? Nope

So which ones ?

Desktop publishing: do people still use dedicated apps for that anymore? AFAIK, people just use Word, which is steadily losing share to Google Docs.

IDEs? here's 13 of them: http://www.hongkiat.com/blog/cloud-ide-developers/

CAD? tinkercad, etc. This one is a classic Christofferson innovator's dilemna. Cloud CAD is the obvious way of doing 3D printing, which can & will move up to eat the big guys.

3D modelling? Clara.io, Lagoa

Mathematical Tools? Good question. A quick google found this, http://www.ms.uky.edu/~statweb/ which is 14 years old... I'm sure if I had a better idea of what to google for, I could find something better...

How many of those services are any good though?

Most likely none.

How many work without an active low-latency internet connection?

Again, most likely none.

How many lock you in to their service?

Probably not all, but I'd seriously doubt "none".

All of them are "good enough" for some users, and have compelling use cases. They aren't feature complete, but they all have small niches where their online nature gives them a compelling advantage over their offline counterparts. From this small niche they can grow "innovator's dilemma" style to be competitive with their desktop competitors.

Google Docs is the most obvious example of this -- I don't know anybody who uses Microsoft Word anymore. Obviously I'm in a little bit of a bubble that way, but they've certainly got a large enough user base to grow from.

Google Docs works while offline, and the same techniques could work on the competitors I mentioned.

In terms of lock-in, all of the services I listed are in the "underdog" phase of growth, where lock-in does more harm than good -- a good interop story with their competitors is more important than the benefits of lock-in. Of course, if they ever reach dominant status this might change, but at the moment, all of the ones I'm familiar with have good & prominent import & export options.

I use LibreOffice, and most people at work use either that or MS Word. At home, it's a mix of LibreOffice and OpenOffice.

My old job used to use Google Docs, but I always felt like I was working against it rather than with it, and I know I wasn't the only one - still a lot of .docx, xlsx, odt, etc being sent around and on fileshares despite the policy being to hand over sensitive company data to google.

You're only looking at highly specialized apps which require a complex UI, so they tend to be more resistant to moving to the web. For the vast majority of apps, which you are ignoring, the UI is just a bunch of forms to fill in. New development of that type of app has almost completely moved to the web.

As well as those mentioned by the other reply:

- Email

- Instant messaging

- Media playing

- Accounting

Mobile has all kinds of problems:

1. Battery life is a joke and that's not going to change.

2. Tiny screen is still tiny. Big screen is more useful.

3. Finger is not the best pointing device.

4. 90% of mobile is games. The market has spoken, smartphone is a toy. Only people that can't afford Internet access (kids & 3rd world) use their phone for everything.

5. Privacy/security is bad from so many angles. Still waiting for an Ubuntu phone.

6. Android is a Linux OS that blocks Linux applications--what BS. If you work at Google, shame on all of you.

7. Processing power is weak. As we're seeing with cryptocoins, don't forget the value of raw computational power.

8. Limiting keyboard. Pressing alt-function to get something simple like an "=" sign is a huge waste of time.

(1) 1.5h used to be good enough for a laptop, not that long ago.

(2) http://en.wikipedia.org/wiki/Mobile_High-Definition_Link

(3) http://www.amazon.com/Logitech-Tablet-Mouse-Android-910-0026...

(6) No, it doesn't. I'm running ssh, git-annex and other Linux applications on my Nexus 7.

I won't disagree about the privacy problems and poor processing power, but I doubt those are real impediments for mobile devices to become the standard, even if it means I'll keep my desktop.

(EDIT: Of course, the "death" of the PC is a silly idea. But I can see it become a professional tool with a small hobbyist market, instead of the consumer market we have now.)

As far as ssh, I bet it was already installed on the device when you got it. Can you uninstall ssh and install it again without rooting Android? If you downloaded some ssh "app" from Play you probably just downloaded a cute interface to the real ssh--which must place nice with the Google proprietary JVM in order to work. Does the real JVM work on Android? Why not? There was a presentation posted on HN recently from a Python developer explaining why he can't code Python apps for Android. He explained it pretty well--you must use Eclipse, you must use the Google Virtual Machine. If you do a few searches for "code Android app" the only way is to download Eclipse and compile through the Google Virtual Machine. I'd rather pluck out my eyeballs.

No, ssh wasn't installed, yes, I can install it again without rooting Android, it's a matter of copying the binary to a directory with executable permissions and running it using at terminal.

f you downloaded some ssh "app" from Play you probably just downloaded a cute interface to the real ssh--which must place nice with the Google proprietary JVM in order to work.

I didn't need an SSH app, and no, Dalvik is not proprietary. Also, you don't need to get apps from Play, you can just download them from the web or use another repository.

Does the real JVM work on Android? Why not?

For the same reason it doesn't work on Glendix - nobody ported it to that userspace.

There was a presentation posted on HN recently from a Python developer explaining why he can't code Python apps for Android. He explained it pretty well--you must use Eclipse, you must use the Google Virtual Machine. If you do a few searches for "code Android app" the only way is to download Eclipse and compile through the Google Virtual Machine.

No, you don't. You just need to cross-compile it to ARM, copy it and run: http://benno.id.au/blog/2007/11/13/android-native-apps

And this is nothing new: if you take notice, the post is from 2007.

Joey Hess writes git-annex in freaking Haskell, compiles the binary using Cabal and I run it just fine on my non-rooted Nexus.

That's fascinating, especially the C compiler, maybe I'll try it. But how to get C applications into Play?

Still the "Allow UNSAFE Binaries?" setting is BS "WARNING: You are about to open your device to hackers! Are you sure? Y / N." You were more patient than I was. I downloaded F-Droid, researched the APK format, almost got an Android emulator working in Debian, even looked at "PhoneGap" yuck. Call me crazy but I prefer Debian over Android. Millions of people rooting their phone I think proves my point--people are willing to risk destroying their device to get root access. Maybe Android would be better if Google had not sold out to the carriers? I don't know.

None of that matters.

Go read "The Innovator's Dilemma".

There's no reason why native apps and the mobile web can't peacefully coexist and thrive. There isn't going to be a "winner." It's not a zero-sum game. The future will be more of what we have today: more apps, more websites, more devices. But nothing is going to die, or be killed off by the other.

I agree. Hence why I said I didn't think Apps were going anywhere either.

Apps have a lot of really cool strengths. But killing the web- that's something people who have trouble seeing long term trends say.

Fighting for mindshare with click-baity headlines ;-)